NOAA (National Oceanic and Atmospheric Administration) satellite images have been widely used in environmental and land cover monitoring. Before delivering to the consumers, NOAA images need to be accurately transformed from image into map. This paper proposes a geometric correction method that corrects the errors caused by this transformation. In this method, the errors are corrected in image coordinate system before the image is transformed into map coordinate system. First, the elevation errors are corrected based on an elevation database. This elevation database is also used to divide data into flat and rough blocks. Next, GCP (Ground Control Point) template matching is applied to specify the residual errors for the blocks that match GCP templates. In order to increase the number of GCP templates, more GCP templates are generated based on the feature of the coastline. The NOAA images from various receiving stations are also referred to make the matching result better. From the residual errors of the blocks which match GCP templates, the residual errors of other flat and rough blocks are calculated by using affine and Radial Basis Function transform respectively. Finally, data is transformed from image into map. With the proposed method, residual errors are corrected more accurately; consequently, the correction results are considerably improved.